A comparison of continuous and discrete foreign gas VA/Q distributions

Estimates of ventilation-perfusion ratio (VA/Q) distributions from foreign gas infusions have utilized a discrete lung model that constrains all alveoli to have 1 of 50 specified VA/Q. In contrast, distribution estimates based on a continuous model allow alveoli to have any VA/Q and can be derived directly from foreign gas data or by transforming distributions derived using the discrete model. We have compared analytically and empirically the properties of distributions based on these two models. Unlike the discrete model, the shape of distributions obtained from the continuous model are insensitive to changes in the VA/Q used in their calculation. Perfusion distributions found from the discrete model using VA/Q values evenly spaced on a logarithmic scale resemble ventilation distributions derived from the continuous model. In contrast, by use of VA/Q evenly spaced on a linear scale, perfusion distributions derived from either model have a similar shape. We concluded that physiological inferences derived from distribution shapes may vary according to the model employed. However, those based upon cumulative perfusion or distribution moments are invariant.